Efficacy of alternative or adjunctive measures to conventional treatment of peri-implant mucositis and peri-implantitis: a systematic review and meta-analysis

In patients with peri-implant mucositis and peri-implantitis, what is the efficacy of nonsurgical (i.e. referring to peri-implant mucositis and peri-implantitis) and surgical (i.e. referring to peri-implantitis) treatments with alternative or adjunctive measures on changing signs of inflammation compared with conventional nonsurgical (i.e. mechanical/ultrasonic debridement) and surgical (i.e. open flap debridement) treatments alone? After electronic database and hand search, a total of 40 publications (reporting on 32 studies) were finally considered for the qualitative and quantitative assessment. The weighted mean changes (WM)/ and WM differences (WMD) were estimated for bleeding on probing scores (BOP) and probing pocket depths (PD) (random effect model). Peri-implant mucositis: WMD in BOP and PD reductions amounted to −8.16 % [SE = 4.61] and −0.15 mm [SE = 0.13], not favouring adjunctive antiseptics/antibiotics (local and systemic) over control measures (p > 0.05). Peri-implantitis (nonsurgical): WMD in BOP scores amounted to −23.12 % [SE = 4.81] and −16.53 % [SE = 4.41], favouring alternative measures (glycine powder air polishing, Er:YAG laser) for plaque removal and adjunctive local antibiotics over control measures (p < 0.001), respectively. Peri-implantitis (surgical): WMD in BOP and PD reductions did not favour alternative over control measures for surface decontamination. WM reductions following open flap surgery (±resective therapy) and adjunctive augmentative therapy amounted to 34.81 and 50.73 % for BOP and 1.75 and 2.20 mm for PD, respectively. While mechanical debridement alone was found to be effective for the management of peri-implant mucositis, alternative/adjunctive measures may improve the efficacy over/of conventional nonsurgical treatments at peri-implantitis sites. Adjunctive resective and/or augmentative measures are promising; however, their beneficial effect on the clinical outcome of surgical treatments needs to be further investigated.


Study inclusion and exclusion criteria
During the first stage of study selection, the titles and abstracts were screened and evaluated according to the following inclusion criteria: 1) English language. 2) Prospective randomized controlled (RCT), or nonrandomized controlled (CCT) studies (split-mouth or parallel group designs) in humans comparing alternative (i.e. for biofilm removal) or adjunctive measures (i.e. for biofilm removal, or adjunctive antiseptic/antibiotic therapy, or regenerative/ resective approaches) to conventional nonsurgical (i.e. mechanical/ultrasonic debridement) or surgical (i.e. open flap debridement) treatments. 3) Data on the clinical changes in mucosal inflammation (i.e. bleeding scores) and probing pocket depths (PD) following nonsurgical (referring to peri-implant mucositis and peri-implantitis) or surgical (referring to peri-implantitis) treatments in respective groups.
At the second stage of selection, all full-text articles identified during the first stage were acquired. During this procedure, the pre-selected publications were evaluated according to the following exclusion criteria: 1) Inclusion of less than five patients 2) Inadequate case definition [13] 3) Lack of clinical data on the changes in mucosal inflammation and PD

Quality assessment of selected studies
A quality assessment of all selected full-text articles was performed according to the Cochrane collaborations tool for assessing risk of bias (low, high, unclear) including the following domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, and incomplete outcome data [14]. Quality assessment was performed in two different phases. In particular, during phase I, quality assessment was based on the published full-text article performed independently by both authors. In phase II, disagreements were resolved by discussion.

Data extraction and method of analysis
A data extraction template was generated and based on the study design, population, case definition, observation period, interventions, comparisons, and primary and secondary outcomes as well as the study quality. For data analysis, the bleeding index (BI)/bleeding on probing (BOP) and PD scores after respective healing periods were defined as primary outcomes. Secondary outcomes included changes in marginal bone levels, immunological/ microbiological parameters and the resolution of periimplant diseases, based on case definitions used in respective publications. RCTs and CCTs not implementing appropriate control measures but reporting on changes in primary outcomes were used for the estimation of the overall efficacy of treatments. Heterogeneity between RCTs, meta-analysis (i.e. weighted mean changes/differences and 95 % confidence intervals, random effect model to account for potential methodological differences between studies), forest plots and publication bias (Egger's regression to quantify the bias captured by funnel plots) were assessed using a commercially available software program (Comprehensive Meta-Analysis V2, Biostat, Englewood, NJ 07631 USA).

Study selection
A total of 368 potentially relevant titles and abstracts were found during the electronic and manual search. During the first stage of study selection, 319 publications were excluded based on title and abstract. For the second phase, the complete full-text articles of the remaining 49 publications were thoroughly evaluated. A total of 19 papers had to be excluded at this stage because they did not fulfil the inclusion criteria of the present systematic review (Table 1).
Finally, a total of 40 publications (reporting on 32 studies) were considered for the qualitative and quantitative assessment (Fig. 1).

Quality and risk of bias assessment of selected studies
The review author's judgement about each risk of bias item for each included RCT is presented in Table 2. In particular, the percentages across all included studies for high, low and unclear risk of bias items were 34.1, 54.8 and 11.1 %, respectively ( Table 2).

Subdivision of selected studies
All selected publications were subdivided according to differences in the treatment protocol: o Nonsurgical treatment of peri-implant mucositis-alternative or adjunctive measures for biofilm removal (2 RCTs and 1 CCT) ( Table 3) o Nonsurgical treatment of peri-implant mucositis-adjunctive antispectic therapy (3 RCTs) (Table 4) o Nonsurgical treatment of peri-implant mucositis-adjunctive antibiotic therapy (2 RCTs) ( Table 5). o Nonsurgical treatment of peri-implantitis-alternative measures for biofilm removal (6 RCTs) ( Table 6) o Nonsurgical treatment of peri-implantitis-adjunctive antiseptic therapy (1 RCT) ( Table 7) o Nonsurgical treatment of peri-implantitis-adjunctive antibiotic therapy (4 RCTs) ( Table 8) o Surgical treatment of peri-implantitis-alternative measures for surface decontamination (3 RCTs and 1 CCT) ( Table 9) o Surgical treatment of peri-implantitis-adjunctive resective therapy (1 RCT) (Table 10) o Surgical treatment of peri-implantitis-adjunctive augmentative therapy (4 RCTs, 4 CCTs) ( Table 11) Nonsurgical treatment of peri-implant mucositis The case definitions markedly differed among the studies investigated. While three studies considered mucosal inflammation in the absence of radiographic bone loss [15][16][17], four studies also accepted a bone resorption of up to 3 mm for defining peri-implant mucositis [18][19][20][21]. Moreover, these studies used several clinical parameters to assess mucosal inflammation, employed various oral hygiene instructions (OHI) and defined different intervals for maintenance care. Despite significant improvements in all of the clinical parameters investigated, test (i.e. alternative or adjunctive methods for biofilm removal, adjunctive antiseptic therapy, or adjunctive antibiotic therapy) and control treatments were commonly associated with residual gingival index (GI), BI, and/or BOP scores at 3 to 12 months after therapy (Tables 3, 4

and 5).
Alternative or adjunctive measures for biofilm removal Three studies reported on alternative or adjunctive measures for biofilm removal (Table 3).
In particular, one RCT and one CCT compared the clinical efficacy of adjunctive air polishing (glycine powder) to OHI and mechanical debridement using either ultrasonic scalers or hand instruments [15,18]. Both test and control groups were associated with significant improvements in mean BI and PD scores after therapy. When evaluating absolute values at 6 months, mean BI and PD scores were significantly lower following adjunctive air polishing to teflon curettes [18]. One RCT compared a repeated (3 and 6 months) monotherapy using an air abrasive device to ultrasonic scaling. After a healing period of 12 months, both groups revealed comparable BOP reductions and frequencies of diseased sites [21] (Table 3).
In particular, one RCT assessed the adjunctive application of phosphoric acid gel to carbon curettes and rubber cup polishing, which was provided once every month in both groups. At 5 months, test sites revealed a significantly higher reduction in mean gingival index (GI) (modified, but not specified) and colony-forming units when compared with control sites, respectively [22]. Porras et al. [20] compared OHI + mechanical debridement with and without local pocket irrigation using chlorhexidine digluconate (CHX) + topical CHX gel application + CHX mouthwash (twice for 10 days). At 3 months, mean mucosal bleeding (mBI), BOP scores and microbiological parameters did not significantly differ between test and control groups. However, the test group revealed a significantly higher change in mean PD scores [20]. In another RCT, topical CHX gel application + full mouth disinfection + CHX mouthrinse (2×/day) and tonsil spraying (1×/day) for 14 days was compared with OHI + mechanical debridement (plastic scaler + polyetheretherketone-coated ultrasonic instruments) + full mouth scaling alone. While both treatment procedures were associated with significant PD reductions at 8 months, the BOP scores did not significantly differ to baseline in both groups [17] (Table 4). Adjunctive antibiotic therapy Two studies reported on the adjunctive antibiotic (i.e. local or systemic) therapy to OHI and mechanical debridement [16,19] ( Table 5).
In particular, one RCT compared the adjunctive local delivery of tetracycline HCl (25 %) fibres for 10 days (test) to mechanical debridement alone (control). A complete or partial fibre loss was noted at three implants after 7 to 10 following application. While test sites were associated with marked BOP improvements, control sites revealed a further increase of BOP scores at 3 months [16]. In another RCT, a total of 45 patients were randomly allocated to either OHI + mechanical debridement (titanium curettes + rubber polishing) + systemic antibiotic medication (Azithromycin® 500 mg day 1 and 250 mg days 2-4) (test), or OHI + mechanical debridement alone (control). The subject-based, perprotocol analysis at 6 months did not reveal any significant differences between test and control groups for all clinical and microbiological parameters investigated [19] ( Table 5).
Despite significant improvements in all of the clinical and microbiological parameters investigated, test (i.e. alternative methods for biofilm removal, adjunctive antiseptic therapy, or adjunctive antibiotic therapy) and control treatments were commonly associated with residual BI and BOP scores at 3 to 12 months after therapy (Tables 6, 7 and 8).
Alternative measures for biofilm removal Six RCTs (corresponding to 7 publications) reported on the efficacy of alternative measures for biofilm removal (Table 6). In particular, two studies employed the same type of an ultrasonic device used with a hydroxyapatite fluid polish [28,31], while two studies reported on erbium-doped yttrium aluminum garnet (Er:YAG) laser monotherapy [24,25] and two publications on glycine powder air polishing [27,33]. One study compared Er:YAG laser monotherapy versus air polishing [30].
At 3 months after therapy, nonsurgical ultrasonic debridement was associated with a reduction in mean BOP scores, whereas these values further increased at control sites (i.e. carbon fibre curettes). However, these differences, as well as those noted for mean PD and radiographic bone level changes, did not reach statistical significance between groups [28]. Similarly, when comparing ultrasonic scaling with mechanical debridement using titanium curettes, Renvert et al. [31] also failed to identify any significant between group differences in mean BI and PD reductions at 6 months. Furthermore, both procedures did not reduce bacterial load [35] (Table 6).
In two RCTs, the efficacy of Er:YAG laser monotherapy was compared to that of mechanical debridement using carbon fibre curettes + adjunctive local antiseptic CHX irrigation/application [24,25]. After 6 months of healing, Er:YAG laser application was associated with significantly lower mean BOP scores than the control treatment. However, these improvements failed to reach statistical significance at 12 months, particularly at advanced sites [24,25] (Table 6). In one RCT, glycine powder air polishing resulted in a significantly higher reduction of mean BOP scores at 3, 6, and 12 months        when compared with mechanical debridement + local antiseptic therapy using CHX. The application of this specific air abrasive device was not associated with any emphysema formation or complications in peri-implant wound healing [27,33]. At more advanced sites, Er:YAG laser monotherapy and glycine powder air polishing resulted in comparable BOP/PD reductions and crestal bone level changes, but failed to reduce bacterial load [36] (Table 6).
Adjunctive antiseptic therapy One multicenter RCT reported on the adjunctive antispective therapy to ultrasonic debridement (Table 7). In particular, a CHX containing matrix was repeatedly applied at 2, 4, 6, 8, 12 and 18 weeks until PD was reduced to ≤5 mm. At 6 months, CHX chips resulted in a significantly higher PD reduction than the placebo chips [23].
Adjunctive antibiotic therapy Three RCTs reported on the adjunctive local antibiotic therapy to mechanical debridement (Table 8). In particular, minocycline microspheres were either applied once [32] at baseline or repeatedly [29] at 30 and 90 days and compared with local antispectic therapy using CHX gel (1.0 %). At 12 months, minocycline was associated with significantly higher BOP (refers to a repeated application) and PD (refers to a single application) reductions when compared with the control group. The radiographic (refers to a repeated application) and microbiological analyses failed to reveal any significant differences between both groups. Similar clinical outcomes were also noted when doxycycline hyclate was used as an adjunct to mechanical debridement [26].
In one RCT, adjunctive local antibiotic therapy (minocycline microspheres) was compared to adjunctive antimicrobial photodynamic therapy. At 12 months, both test and control groups were associated with significant but comparable clinical, microbiological and immunological improvements [34,37] (Table 8).
Alternative measures for surface decontamination In one CCT, Deppe et al. [40] assessed the clinical efficacy of carbon dioxide laser decontamination used as an adjunct to resective flap surgery + air polishing (control). While the test treatment improved the clinical outcomes over the control measure at 4 months, mean SBI and PD values were comparable in both groups at about 5 years (Table 9). In further two studies, De Waal et al. [38,39] employed an open flap debridement using gauzes soaked in sterile saline + bone re-contouring + apical flap re-positioning and compared one test (0.12 % CHX + 0.05 % cetylpyridinium chloride) and two control (placebo solution or 2.0 % CHX) measures for surface decontamination. At 12 months, the test and both control procedures were associated with marked but comparable reductions in mean BOP and PD scores, respectively. Furthermore, between group comparisons    Open flap surgery + debridement + decontamination using an Er:YAG laser device (cone-shaped glass fibre tip) at 11.4 J/cm 2 implantoplasty at bucally and supracrestally exposed implant parts Open flap surgery + Mechanical debridement (plastic curettes) + decontamination (cotton pellets soaked in saline) implantoplasty at bucally and supracrestally exposed implant parts   Subgroup analysis of n = 13 implants failed to reveal any significant differences in mean marginal bone loss after therapy [38,39]. Similarly, Papadopoulos et al. [41] also failed to reveal any significant clinical improvements in mean BOP and PD scores at 6 months, when a 980-nm diode laser was used as an adjunct to mechanical open flap debridement (Table 9).
Adjunctive resective therapy One study assessed the clinical efficacy of an implantoplasty (diamond/arkansas burs + silicone polishers) when used as an adjunct to open flap debridement + bone re-contouring + apical flap re-positioning [42]. At 24 months, all patients from the control group had to be discontinued from the study due to persistent active signs of peri-implant inflammation. This was associated with elevated mBI and PD scores when compared with the test group. On the contrary, resective therapy resulted in significantly higher mean mucosal recessions (1.64 ± 1.29 vs. 2.3 ± 1.45 mm) but no pseudopocket formation [42], while test sites were associated with stable radiographic bone levels at 3 years, the interproximal bone loss at control sites amounted to 1.45-1.54 mm [57] ( Adjunctive augmentative therapy The clinical efficacy of adjunctive augmentative therapy to open flap debridement (titanium curettes + conditioning using 24 % ethylenediaminetetraacetic acid gel + submerged healing for 6 months) was merely assessed in one study [55]. Notably, 12/16 control and 13/16 test sites revealed a premature exposure during the submerged healing phase of 6 months. At 12 months after therapy, the application of porous titanium granules to the intrabony defect components resulted in a significantly higher percentage of radiographic defect fill when compared with open flap surgery alone. Moreover, the test group was associated with an increase in implant stability quotient, whereas these values further decreased at control sites. However, both groups revealed comparable PD reductions and only minor improvements in mean BOP scores [55]. The immunological analysis did not reveal any significant between differences in the reduction of MMP-8 levels or bone level markers [58] ( Table 11). Four RCTs and four CCTs compared different augmentation protocols employing various methods for surface decontamination, bone fillers (i.e. alloplastic, xenogenic, autogenous) and barrier membranes (synthetic, native collagen) over a period of up to 5 years [40,[43][44][45][46][47][48][49][50][51][52][53][54][55]. The majority of these studies considered PD and BOP reductions as primary outcomes but also reported on radiographic defect fill (Table 11). WM BOP [43,49,51,53,55] and PD [40,[43][44][45][46][47][48][49][50][51][52][53][54][55]  , respectively. The outcomes of therapy was mainly influenced by the type of bone filler (i.e. a slowly resorbing bovine-derived mineral was superior to autogenous bone and an alloplastic material), defect characteristics (i.e. circumferential-type defects were superior to dehiscence-type defects) and implant surface characteristics (i.e. moderately rough surfaces were superior to rough surfaces) ( Table 11).

Meta-analysis
Meta-analysis to estimate the weighted mean difference (WMD) between test and control treatments was conducted on RCTs reporting on similar assessments of either absolute BOP or PD scores.
Egger's linear regression method revealed symmetrical plots for changes in BOP (p = 0.51) and PD (p = 0.69) thus suggesting the absence of publication bias.
Egger's linear regression method revealed symmetrical plots for changes in BOP (p = 0.41) and PD (p = 0.39) thus suggesting the absence of any publication bias.
Egger's linear regression method revealed symmetrical plots for changes in BOP (p = 0.17) and PD (p = 0.07) thus suggesting the absence of any publication bias.
Surgical treatment of peri-implantitis -alternative measures for surface decontamination Based on two studies [38,39], WMD in BOP and PD scores between test and control groups amounted to 5.  (Fig. 4a, b).

Discussion
The present systematic review and meta-analysis was conducted to address the following focused question: "In patients with peri-implant mucositis and periimplantitis, what is the efficacy of nonsurgical (i.e. referring to peri-implant mucositis and peri-implantitis) and surgical (i.e. referring to peri-implantitis) treatments with alternative or adjunctive measures on changing signs of inflammation compared with conventional nonsurgical and surgical treatments alone?". Basically, the literature search revealed that only a few studies considered appropriate test and control treatments needed to address the aforementioned focused question. In particular, this was true for 8 (7 RCTs and 1 CCT) [15][16][17][18][19][20][21][22] and 9 (9 RCTs) [23-26, 28, 29, 31-33] studies reporting on the nonsurgical treatment of periimplant mucositis and peri-implantitis, as well as 5 RCTs [38,39,41,42,55] reporting on the surgical treatment of peri-implantitis. In addition, 5 RCTs and 4 CCTs not implementing appropriate control measures (i.e. mechanical/ultrasonic debridement or open flap debridement alone) but reporting on changes in primary outcomes were included for the estimation of the overall efficacy (referring to WM changes in BOP and PD scores) of nonsurgical [30,34] and surgical [40,43,44,46,48,52,53] treatments of peri-implantitis. Moreover, it must be emphasized that the percentage across all included studies for high risk of bias items was 34.1 %, thus pointing to a need to improve the quality of reporting in future studies.
Within these limitations, the current data synthesis revealed that for the nonsurgical treatment of periimplant mucositis, WMD in BOP [16,17,19,22] and PD [15,17,19,20,22] scores amounted to −8.16 % and −0.15 mm, not favouring local antiseptic or antibiotic (i.e. local and systemic) therapy as an adjunct to mechanical debridement alone. Basically, these data corroborate the findings of a recent systematic review and meta-analysis, also indicating that adjunctive therapy may not improve the efficacy of professionally administered plaque removal in reducing BOP (i.e. local antiseptic or local/systemic antibiotics), GI and PD (i.e. local antiseptics, systemic antibiotics, air abrasive device) scores at mucositis sites [9]. When considering the present narrative data synthesis on the adjunctive [15] or alternative use of glycine powder air polishing [21], it was also noted that this device did not reveal any major improvements in BI/BOP scores or disease resolution over the respective control measures. In this context, it must be emphasized that BOP is the key parameter for the diagnosis of peri-implant mucositis [1], and the "resolution of peri-implant mucosal inflammation as evidenced by the absence of BOP" was the suggested endpoint following nonsurgical treatment of mucositis lesions [13]. All these data, taken together with the present findings support the view that OHI and mechanical debridement with or without polishing tools may be defined as a current standard of care for the management of peri-implant mucositis [6].
Since the suggested endpoint following nonsurgical treatment of peri-implantitis is a "composite outcome of disease resolution including the absence of deep PD with bleeding and suppuration" [13], one has to critically emphasize the limited efficacy at "deep sites". In particular, several studies reported on increasing BOP scores between 3 and 12 months following nonsurgical treatment of "severe" peri-implantitis sites using either mechanical debridement, adjunctive aPDT, Er:YAG laser monotherapy or glycine powder air polishing. The efficacy of all treatment procedures investigated was higher at "moderately" deep sites [24,27,33,59]. In this context, however, one also has to realize that PD scores at implant sites may be influenced by a variety of different local factors, including the soft tissue thickness, vertical implant positioning, or a specific design of the implant-abutment connection (e.g. platform-switching). Accordingly, the classification of a "deep" pocket needs to be accomplished on an individual basis and disease severity should also consider "proportions of affected implants per patient in the presence of multiple implants" [6].
When further analysing the present results, it was also noted that nonsurgical treatment of peri-implantitis commonly failed to result in major microbiological improvements [32,[34][35][36][37], thus potentially explaining the frequency of residual BOP scores at respective sites.
At the time being, there is a lack of clinical studies aimed at comparing the efficacy of nonsurgical and surgical treatments of peri-implantitis. However, a preclinical study employing the ligature model has indicated that open flap debridement was associated with significant histological improvements in osseous defect fill and establishment of a new bone-to-implant when compared with nonsurgical treatments. The latter outcome was mainly influenced by the method of surface decontamination [60]. Accordingly, a "proven method of decontaminating the implant surface" has been defined as a critical component in surgical therapy [13]. However, the present qualitative and quantitative analysis has indicated that the clinical, radiographical and microbiological outcomes following either open flap debridement or surgical augmentative therapy were not influenced by the decontamination protocol, including chemical or photothermal (i.e. carbon dioxide, diode-or Er:YAG laser radiation) approaches [38-41, 49, 50, 52]. Moreover, two RCTs [38,39] reported on an additional bone loss at 6 and 12 months after open flap debridement, thus indicating that disease resolution (i.e. "absence of deep probing pocket depths with bleeding and suppuration and no additional bone loss") [13] was commonly not achieved.
The present data synthesis also revealed a lack of RCTs/CCTs implementing appropriate test and control groups to assess the efficacy of adjunctive resective or augmentative measures over open flap debridement alone. The available studies have indicated that resective surgery (i.e. apical re-positioned flap + bone contouring) + implantoplasty was more effective in obtaining and maintaining disease resolution over resective surgery alone [42,57]. In contrast, surgical augmentative therapy of the intrabony defect component using porous titanium granules was associated with a significantly higher radiographic defect fill, but failed to improve a reduction in mean BOP and PD scores over the control treatment [55]. When considering the estimated WM BOP (50.73 vs. 34.81 %) [38,39,43,49,51,53,55] and PD (2.20 vs. 1.75 mm) [38][39][40][43][44][45][46][47][48][49][50][51][52][53][54][55] reductions, the clinical outcomes obtained following adjunctive augmentative therapy tended to be improved when compared with surgical measures alone. However, it has to be realized that for the data synthesis in the latter group, surgical procedures with and without soft tissue resection were combined, and therefore, the interpretation of the overall performance of surgical measures without augmentative measures on PD reductions is difficult. Moreover, obvious variations in the surgical procedures, including different decontamination protocols, administration of prophylactic systemic antibiotics, and modes of healing (i.e. open vs. submerged) may not allow for a direct comparison of these estimated outcomes. Basically, the estimated WM BOP and PD reductions corroborate those calculations reported in a recent systematic review on reconstructive procedures for the management of peri-implantitis. When also case series were included in the meta-analysis, these values amounted to 45.8 and 2.97 mm [61]. Furthermore, the present qualitative analysis of the available data on surgical augmentative therapy have indicated, that the outcomes of therapy may be influenced by several local factors, mainly including the physicochemical properties of the bone filler [43,48,51,54], the defect configuration [53], as well as implant surface characteristics [44]. Any beneficial effect of a resorbable synthetic barrier membrane could not be identified [45][46][47]. Nevertheless, the available evidence did not allow for any conclusive statements on the potential superiority of any particular augmentation protocol.
Finally, it must be emphasized that laser therapy, the application of bone grafts and barrier membranes were reported to be associated with the highest cost-effectiveness ratio (i.e. costs and proportions of lost implants) among 11 treatment procedures investigated [62].

Conclusions
While OHI + mechanical debridement alone was found to be effective for the management of peri-implant mucositis, alternative/adjunctive measures may improve the efficacy over/of conventional treatments at peri-implantitis sites. Adjunctive resective and/or augmentative measures are promising; however, their beneficial effect on the clinical outcome of surgical treatments needs to be further investigated.

Competing interests
Frank Schwarz and Jürgen Becker received lecture fees and/or unrestricted research grants from Geistlich Biomaterials, Rooth, Switzerland; Osteology Foundation, Lucerne, Switzerland; EMS, Nyon, Switzerland. Frank Schwarz, Andrea Schmucker and Jürgen Becker declare that they have no competing interests related to this systematic review.
Authors' contributions FS and JB have made substantial contributions to study conception and interpretation of data as well as manuscript preparation. AS and FS were involved in the literature search and data management as well as the statistical analysis. All authors read and approved the final manuscript.